A new round of leaks out of the GPU rumor mill suggests Nvidia's next flagship graphics card, the RTX 5090, could draw as much as 600 watts under full load.
That's a jaw-dropping number when you consider the RTX 4090 already pulls around 450 watts.
If accurate, this isn't just an incremental upgrade.
It's a signal that the company is done pretending power efficiency matters at the top of the stack.
The whispers trace back to supply chain chatter and forum posts from leakers who have been reliable before, though Nvidia hasn't confirmed a thing.
The rumored card is said to pack a GB202 chip, 32GB of GDDR7 memory, and a memory bus that could hit 512 bits.
Translation for anyone who doesn't live in spec sheets: this thing is built to chew through 4K and 8K gaming, AI workloads, and whatever else you throw at it without breaking a sweat.
Here's where it gets interesting for regular people.
A 600-watt card doesn't just change your GPU budget.
It changes your power supply, your case airflow, your electricity bill, and possibly the wiring in older homes.
You'd realistically need a 1,200-watt PSU to leave any headroom, and that's before you factor in a high-end CPU that itself wants 150 watts or more.
The total system draw starts looking like a small space heater.
The pricing picture is murky, but nobody expects mercy.
The RTX 4090 launched at $1,599 and street prices stayed well above that for months.
If the 5090 arrives with a bigger die, exotic GDDR7 memory, and a beefier cooling solution, a $2,000-plus launch price wouldn't shock anyone watching this market.
Nvidia has figured out that enthusiasts and AI researchers will pay almost anything when there's no real competition at the top.
That lack of competition is the part worth sitting with.
AMD has publicly stepped back from chasing the absolute high end, and Intel's Arc line is still finding its footing.
When one company owns the ceiling, it sets the terms.
It's a flex, and the rest of the PC building world gets to react to it.
For most gamers, the smarter move may be waiting to see what the 5080 looks like.
Historically, the second-tier card delivers most of the real-world performance for hundreds less, and the 5090's absurd power envelope could make it a niche purchase even among deep-pocketed builders.
The mid-range card sells to actual humans.
Nvidia is juggling insatiable demand for its AI accelerators, and consumer GPUs share fabs and memory supply with those money-printing chips.
Every wafer that goes into a gaming card is a wafer not going into a data center part.
That dynamic could push the 5090's launch window and its availability in ways nobody outside Nvidia can predict.
What's clear is that the top of the consumer GPU market is drifting away from the average PC owner.
Cards like this are increasingly built for people running local AI models, rendering farms, and 4K sim rigs.
The rest of us watch the keynote, nod at the benchmarks, and quietly buy something sensible.
My take: the 5090 will be a genuinely impressive piece of engineering that almost nobody needs and a small number of people will happily overpay for.
It's how far the flagship tier has drifted from what a normal American gamer can justify spending.
Final Thoughts
Watch the power connector news closely, because that's where the drama always lands.